That's not so funny... Just last week Nat Freidman and Miguel were debating whether Gnome should push java or mono.
To me it's simpler to just push mono. Otherwise you have to deal with half a dozen different incompatible JVMs. With mono there is just one Linux CLI so it's much easier to support. If Sun open sourced their java stuff, that would create a defacto standard or instead they can chose not to and probably we'll all use mono.
I agree that Sun shouldn't have to care about Linux users.
Everyone should get behind IBM's implementation and standardize on that. The current situation is absolutely nuts. Java programs larger than "Hello World" have about a 25% chance of running on a default Linux system.
I used to prefer Java to C++ but it's too difficult to support because the Linux JVMs are all incompatible. I wouldn't distribute a program written in Java at this point. I don't think I'll go back to programming in JAVA unless something really changes. Probably programming with mono will be an easier, better solution in a year or two.
How many children will be saved later on from what we learn with this operation?
This operation is so amazing for its own sake and if they can save the baby as well, that's two birds with one stone! It doesn't get more excelent than that.
>> he is making the tacit assumption that many of these lawsuits have merit, and that much of the liability is real.
He's definately saying that the liability is real but he doesn't care whether the lawsuits have merit or not.
It's like paying the mob and bribing the cops. Dealing with lawsuits is just something you have to do in the OS space that's why it takes a company with a multi-billion dollar market cap to survive.
I do read the article as saying that Microsoft is behind the SCO lawsuit and that he suspects Microsoft has 50 similar lawsuits planned.
"I suspect Microsoft may have 50 or more of these lawsuits in the queue. All of them are not asking for hundreds of millions, but most would be large enough to ruin anything but the largest companies."
It's ambigous. Is he talking about the patent lawsuit against Microsoft or is he talking about the SCO lawsuit? I think the second meaning is more likely but it is ambiguous.
Your post is a pretty good summary... One thing that I noticed:
> Better leave it to the big guys (Sun, MS, SCO...)
SCO is a tiny guy.
RHAT: Market Cap: 3.16B SCOX: Market Cap: 136.35M
One point that Anderer does bring up is that anyone could have made this issue go away just by buying SCO when they had a $10 million market cap. In a pure business world it makes sense to just give in and "pay the mob" so to speak.
It's not hard to find $10 million dollars. RedHat set up a $1 million dollar defense fund. OSDL has their $10 million dollar fund. IBM has spent more than $10 million on litigation already.
It's not a money issue. It has to do with integrety and bride.
Leggett & Platt Inc: Claims they don't remember buying the licenses. Are not interested in buying any SCO licenses.
CA: Claims they bought UnixWare licenses for their customers as part of their $40 million settlement with the Canopy Group. Claims they didn't buy actual licenses.
Questar: Admits they bought a couple licenses for $5000.
These obviously affect both IBM's counter lawsuit and RedHat's lawsuit against SCO. I wouldn't be surprised if Microsoft tells SCO to fold now that their game is up. Probably in a few months.
The slant you suggested is a bit too violent for my taste...
When I take an unbiased view of the situation, I see that SCO is basically trying to trick people into giving them money. SCO doesn't have the copyright assignments from Novell. As far as I know they haven't shown any examples of infringing code. They are still distributing the 2.4 Linux kernel from their website. SCO is going around harassing and threatening Linux customers.
Some things that people do are legal and some things are illegal. SCO is doing the illegal stuff.
If I ran the zoo I would sentence Darl to 3 months in prison. That won't happen, but I think it would be completely reasonable. I think that if Darl was in prison for a few months he might reform and potentially be a better person for it.
Probably the CEO was promised kickbacks from Microsoft for signing up with SCO. That's fine if that's how he wants to make money, but we want our share in return.
Nah... I've done a bit of auditing and it's no big deal to tell someone there is a mistake in the code somewhere. Mostly they just fix the problem or tell me I'm missing something.
I once got into a spat with a junior developer who didn't realize why what he was doing was wrong. In that case I phrased the problem as a question and said, "Won't this cause problems?" instead of saying that it was obviously wrong. We got into a big stupid discusion about whether it would or not and he ended up unconvinced.
It would have been better to just say straight out, "Probably you just did this without thinking but it's going to cause problems. Could you find some fix for that?"
OpenBSD is grouchy because the senior developers are grouchy.
I like the idea of breaking things up from a technical/development/organizational standpoint but I don't like firebird UI.
I like the way the middle click opens a new window in Mozilla, and I don't like tabs. Currently I have 8 browser windows open, my xchat window is open, I have 4 eterm windows open and 4 random windows minimized. That's because I was shutting down for the day and exited a bunch of applications.
My window manager is set up so that I can take advantage of Fitt's Law to switch between windows quickly. Tabs do not take advantage of Fitt's law so they are slow by default, but after you have a lot of windows open they become unmanageable.
Most people do not have a good window manager like mine so switching between windows is slow and confusing. In that case, tabs make a lot of sense.
In some ways, this kind of math is a funny thing to benchmark. Adding and multiplication are just tiny assembly language instructions that are the same no matter what human readable programming language the test originally used. Any differences in the run time is going to be caused by other parts of the language.
Python did pretty badly in the tests. The reason is that in Python it takes a long time to translate a variable name into a memory address (It happens at runtime instead of compile time).
The benchmark code has stuff that basically looks like this:
while foo < maxInt:
i += 1
Adding 1 to i takes no time at all but looking up i take a little time. In C this is going to be a lot faster.
Python did really bad when "i" from the example above was a long compared to when it was a long in C. That's because Python has big number support but in C a long is limited to just 4 bytes.
Python did OK in the trig section because the trig functions are implemented in C. It still suffers because it takes a long time to look up variables though.
In real life, variable look up time is sometimes a factor. However, for programs that I've written getting data from the network, or database was the bottleneck.
Think of it more as easy to access bookmarks with some predefined bookmarks by default.
If you don't like the bookmarks just right click and remove. If you want to bookmark a different directory just drap 'n drop. This is way more useful than predefined short cuts like I have seen before.
You must file SEC forms months or years in advance,
Probably "months" is a little bit more likely than "years".
If you look at the timeline SCO started planning the lawsuit around October 2002, they filed the documents with the SEC in January 2003, and they filed the lawsuit against IBM in March. It's hard to argue that it was pure coincidence that SCO executives started dumping stock right after the lawsuit was announced. And as far as pumping goes SCO is the best.
Back in the day Xfree86 needed to be a corporation to trademark the term "Xfree86" so they created this weird organization with a constitution and everything. There was the board and there was the core. Later another group was added, people who had commit access to the CVS repository, but weren't on the core. Then at the bottom there were regular developers.
The problem is that no one really new what the core does except that it had a private email list. Even people on the core didn't know. (I'm not making this up).
Historically XFree86 has had closed developement. If you wanted to read the developers emails or look at the development code you had to apply and be approved. A couple years ago they openned up the CVS repository to the world. Then earlier this year they openned up all the development email lists.
But once in a while when during code discussions people would say, "Oh that. We discussed on the core email list and we decided it sucked. Case closed." That kind of thing gets annoying.
Some people said that the core email list should be destroyed, but the answer was that, "Why do you care? All the development discussion is on the developers email list." This was probably true in theory if not in real life.
To be on the core you had to be selected after coding for 3 or 4 years. It's not worth it really because as I said, no one knows what the core does and all the real power is held by the people with CVS commit access anyway.
The whole idea of a core group was stupid and pointless. The reason it stuck around for so long was that XFree86 developers are stubborn people. Everyone (often not developers) was telling them to change and have elections and so they said, "Screw you. We'll do whatever we want." Another reason was that some people on the core group weren't active developers and didn't follow the lists closely. They didn't realise how frustrated people were.
I've been getting more and more upset as I write this post thinking about how XFree86 used to be, but I started out just wanting to say that it was a good thing. I believe it is a good thing for XFree86. It's a sign that the project is becoming more transparent and responsive to developers. It takes humility on the part of the core members to give up the extra privileges.
>>Recently I saw a machine with kernel 2.4.something (SuSE 8.2) which ran out of memory and a lockup happened.
You could probably hear the drive grinding away in the background right? That's because the kernel is trying to move stuff around to swap space or back. After a while, the kernel just picks a program and kills it to free up memory. I don't know the algorithm for picking which program to kill, but for me it normally kills mozilla.
Marcelo removed the Out Of Memory killer from 2.4 a couple weeks ago. I'm not sure how it works now.
When I have low memory, I get messages like this: timeoutd: page allocation failure. order:0, mode:0x20 After that the system locks up completely and the NumLock key stops working. When the NumLock key stops that means the kernel is not responding to interrupts.
This test would have been more interesting if there had been failures. Perhaps they could have tried the test on an older version of Linux, or a different operating system.
I have been trying to write some tests of my own recently. So far I have found a filesystem OOPs, a ptrace BUG(), and my system locks up on low memory situations. Probably the lockup is because my ethernet driver allocates memory in the interrupt handler (GFP_ATOMIC) and can't handle the result when there is no memory available.
I need to fix the lock up first of all so the other tests have time to run...
Slashdot Mirror
That's not so funny... Just last week Nat Freidman and Miguel were debating whether Gnome should push java or mono.
To me it's simpler to just push mono. Otherwise you have to deal with half a dozen different incompatible JVMs. With mono there is just one Linux CLI so it's much easier to support. If Sun open sourced their java stuff, that would create a defacto standard or instead they can chose not to and probably we'll all use mono.
I agree that Sun shouldn't have to care about Linux users.
Everyone should get behind IBM's implementation and standardize on that. The current situation is absolutely nuts. Java programs larger than "Hello World" have about a 25% chance of running on a default Linux system.
I used to prefer Java to C++ but it's too difficult to support because the Linux JVMs are all incompatible. I wouldn't distribute a program written in Java at this point. I don't think I'll go back to programming in JAVA unless something really changes. Probably programming with mono will be an easier, better solution in a year or two.
Because slashdot slashdots their pitiful database and no one can get any work done. It happenned a couple times and people got upset.
How many children will be saved later on from what we learn with this operation?
This operation is so amazing for its own sake and if they can save the baby as well, that's two birds with one stone! It doesn't get more excelent than that.
>> Paul Allen funded SCO harrassment case
Paul Allen is a decent guy. Also he doesn't own Vulcan so he's not funding SCO that I know of.
>> he is making the tacit assumption that many of these lawsuits have merit, and that much of the liability is real.
He's definately saying that the liability is real but he doesn't care whether the lawsuits have merit or not.
It's like paying the mob and bribing the cops. Dealing with lawsuits is just something you have to do in the OS space that's why it takes a company with a multi-billion dollar market cap to survive.
I do read the article as saying that Microsoft is behind the SCO lawsuit and that he suspects Microsoft has 50 similar lawsuits planned.
"I suspect Microsoft may have 50 or more of these lawsuits in the queue. All of them are not asking for hundreds of millions, but most would be large enough to ruin anything but the largest companies."
It's ambigous. Is he talking about the patent lawsuit against Microsoft or is he talking about the SCO lawsuit? I think the second meaning is more likely but it is ambiguous.
Your post is a pretty good summary... One thing that I noticed:
> Better leave it to the big guys (Sun, MS, SCO...)
SCO is a tiny guy.
RHAT: Market Cap: 3.16B
SCOX: Market Cap: 136.35M
One point that Anderer does bring up is that anyone could have made this issue go away just by buying SCO when they had a $10 million market cap. In a pure business world it makes sense to just give in and "pay the mob" so to speak.
It's not hard to find $10 million dollars. RedHat set up a $1 million dollar defense fund. OSDL has their $10 million dollar fund. IBM has spent more than $10 million on litigation already.
It's not a money issue. It has to do with integrety and bride.
Leggett & Platt Inc: Claims they don't remember buying the licenses. Are not interested in buying any SCO licenses.
CA: Claims they bought UnixWare licenses for their customers as part of their $40 million settlement with the Canopy Group. Claims they didn't buy actual licenses.
Questar: Admits they bought a couple licenses for $5000.
EV1: Admits they bought a license.
The had just settles a lawsuit with the Canopy Group for $40 million and the IP license was thrown in for free.
These obviously affect both IBM's counter lawsuit and RedHat's lawsuit against SCO. I wouldn't be surprised if Microsoft tells SCO to fold now that their game is up. Probably in a few months.
The slant you suggested is a bit too violent for my taste...
When I take an unbiased view of the situation, I see that SCO is basically trying to trick people into giving them money. SCO doesn't have the copyright assignments from Novell. As far as I know they haven't shown any examples of infringing code. They are still distributing the 2.4 Linux kernel from their website. SCO is going around harassing and threatening Linux customers.
Some things that people do are legal and some things are illegal. SCO is doing the illegal stuff.
If I ran the zoo I would sentence Darl to 3 months in prison. That won't happen, but I think it would be completely reasonable. I think that if Darl was in prison for a few months he might reform and potentially be a better person for it.
You can still download the 2.4 linux kernel without enterring any passwords.
o rk station/current/floppy/install.144
ftp://ftp.sco.com/pub/updates/OpenLinux/3.1.1/W
That's a floppy image with the 2.4 Linux kernel on it.
>>Civil disobedience is not a good business model.
The only company breaking the law is SCO.
Stowell didn't say that EV1 paid $1 million, he said the deal was "worth at least $1 million". That means that 699 * number_of_servers > 1,000,000.
What probably happenned was that Microsoft and EV1 worked out an advertising deal and SCO charged EV1 90% off the regular price.
What needs to happen now is that EV1 needs to donate to the OSDL defense fund or the Linux community needs to boycott EV1.
Probably the CEO was promised kickbacks from Microsoft for signing up with SCO. That's fine if that's how he wants to make money, but we want our share in return.
Everytime Bill Gates looks in the mirror he sees Bill Gates looking back. Also he spends a lot of time with Steve Balmer.
Either of those things mean you can claim to be in a living hell.
Nah... I've done a bit of auditing and it's no big deal to tell someone there is a mistake in the code somewhere. Mostly they just fix the problem or tell me I'm missing something.
I once got into a spat with a junior developer who didn't realize why what he was doing was wrong. In that case I phrased the problem as a question and said, "Won't this cause problems?" instead of saying that it was obviously wrong. We got into a big stupid discusion about whether it would or not and he ended up unconvinced.
It would have been better to just say straight out, "Probably you just did this without thinking but it's going to cause problems. Could you find some fix for that?"
OpenBSD is grouchy because the senior developers are grouchy.
I like the idea of breaking things up from a technical/development/organizational standpoint but I don't like firebird UI.
I like the way the middle click opens a new window in Mozilla, and I don't like tabs. Currently I have 8 browser windows open, my xchat window is open, I have 4 eterm windows open and 4 random windows minimized. That's because I was shutting down for the day and exited a bunch of applications.
My window manager is set up so that I can take advantage of Fitt's Law to switch between windows quickly. Tabs do not take advantage of Fitt's law so they are slow by default, but after you have a lot of windows open they become unmanageable.
Most people do not have a good window manager like mine so switching between windows is slow and confusing. In that case, tabs make a lot of sense.
Python did pretty badly in the tests. The reason is that in Python it takes a long time to translate a variable name into a memory address (It happens at runtime instead of compile time).
The benchmark code has stuff that basically looks like this:
Adding 1 to i takes no time at all but looking up i take a little time. In C this is going to be a lot faster.
Python did really bad when "i" from the example above was a long compared to when it was a long in C. That's because Python has big number support but in C a long is limited to just 4 bytes.
Python did OK in the trig section because the trig functions are implemented in C. It still suffers because it takes a long time to look up variables though.
In real life, variable look up time is sometimes a factor. However, for programs that I've written getting data from the network, or database was the bottleneck.
Think of it more as easy to access bookmarks with some predefined bookmarks by default.
If you don't like the bookmarks just right click and remove. If you want to bookmark a different directory just drap 'n drop. This is way more useful than predefined short cuts like I have seen before.
You must file SEC forms months or years in advance,
Probably "months" is a little bit more likely than "years".
If you look at the timeline SCO started planning the lawsuit around October 2002, they filed the documents with the SEC in January 2003, and they filed the lawsuit against IBM in March. It's hard to argue that it was pure coincidence that SCO executives started dumping stock right after the lawsuit was announced. And as far as pumping goes SCO is the best.
Back in the day Xfree86 needed to be a corporation to trademark the term "Xfree86" so they created this weird organization with a constitution and everything. There was the board and there was the core. Later another group was added, people who had commit access to the CVS repository, but weren't on the core. Then at the bottom there were regular developers.
The problem is that no one really new what the core does except that it had a private email list. Even people on the core didn't know. (I'm not making this up).
Historically XFree86 has had closed developement. If you wanted to read the developers emails or look at the development code you had to apply and be approved. A couple years ago they openned up the CVS repository to the world. Then earlier this year they openned up all the development email lists.
But once in a while when during code discussions people would say, "Oh that. We discussed on the core email list and we decided it sucked. Case closed." That kind of thing gets annoying.
Some people said that the core email list should be destroyed, but the answer was that, "Why do you care? All the development discussion is on the developers email list." This was probably true in theory if not in real life.
To be on the core you had to be selected after coding for 3 or 4 years. It's not worth it really because as I said, no one knows what the core does and all the real power is held by the people with CVS commit access anyway.
The whole idea of a core group was stupid and pointless. The reason it stuck around for so long was that XFree86 developers are stubborn people. Everyone (often not developers) was telling them to change and have elections and so they said, "Screw you. We'll do whatever we want." Another reason was that some people on the core group weren't active developers and didn't follow the lists closely. They didn't realise how frustrated people were.
I've been getting more and more upset as I write this post thinking about how XFree86 used to be, but I started out just wanting to say that it was a good thing. I believe it is a good thing for XFree86. It's a sign that the project is becoming more transparent and responsive to developers. It takes humility on the part of the core members to give up the extra privileges.
This is a good thing for everyone.
And people said that Windows wasn't as portable as Linux or NetBSD. What do you have to say now, punks?!
>>Recently I saw a machine with kernel 2.4.something (SuSE 8.2) which ran out of memory and a lockup happened.
You could probably hear the drive grinding away in the background right? That's because the kernel is trying to move stuff around to swap space or back. After a while, the kernel just picks a program and kills it to free up memory. I don't know the algorithm for picking which program to kill, but for me it normally kills mozilla.
Marcelo removed the Out Of Memory killer from 2.4 a couple weeks ago. I'm not sure how it works now.
When I have low memory, I get messages like this:
timeoutd: page allocation failure. order:0, mode:0x20
After that the system locks up completely and the NumLock key stops working. When the NumLock key stops that means the kernel is not responding to interrupts.
This test would have been more interesting if there had been failures. Perhaps they could have tried the test on an older version of Linux, or a different operating system.
I have been trying to write some tests of my own recently. So far I have found a filesystem OOPs, a ptrace BUG(), and my system locks up on low memory situations. Probably the lockup is because my ethernet driver allocates memory in the interrupt handler (GFP_ATOMIC) and can't handle the result when there is no memory available.
I need to fix the lock up first of all so the other tests have time to run...